The ⅔ trellis coding used in 8VSB DTV broadcasting already restricts the 8VSB symbol alphabet, which includes normalized modulation levels of −7, −5, −3, −1, +1, +3, +5 and +7. 8VSB without ⅔ trellis coding would have a symbol-error-distance distance of two normalized modulation levels. The ⅔ trellis coding can be exploited to increase the symbol-error-distance in effect, so an error of at least four normalized modulation levels is required to cause an error in decision. A signal-to-noise ratio (SNR) of at least 14.5 dB is required for the trellis coded 8 VSB DTV signal to exceed the threshold of visibility (TOV) of errors in the received DTV signal.
Certain ways of further restricting the 8VSB symbol alphabet to increase robustness of over-the-air transmission were specifically described in U.S. patent application Ser. No. 10/078,933 titled “ATSC DIGITAL TELEVISION SYSTEM” and filed 19 Feb. 2002 by D. Birru, V. R. Gaddam and M. Ghosh. The 5 Dec. 2002 publication number is 2002-0181581, and the application is assigned to Koninklijke Philips Electronics, N.V. This Philips application describes hierarchical 8VSB modulation, trellis-coded 4VSB modulation, and pseudo-2VSB (P-2VSB) modulation.
U.S. patent application Ser. No. 10/078,933 teaches that P-2VSB modulation can be generated by choosing the Y2 bit in each two-bit symbol supplied to the ⅔ trellis encoder in an 8VSB DTV transmitter to be the same as the X1 bit. This restricts the 8VSB symbol alphabet to consist of only those symbols corresponding to normalized modulation levels of −7, −5, +5 and +7. In the resulting pseudo-2VSB or P-2VSB modulation a symbol-error-distance of at least ten normalized modulation levels is required to cause an error in decision between the 00 and the 11 two-bit symbols. Presuming the transmitted data to be random, P-2VSB modulation has a higher ratio of average power to peak power than 8VSB modulation. When P-2VSB signal is time-division multiplexed with 8VSB signal, there is a need to cut back on DTV transmitter modulation levels to stay within the bounds of average power permitted to the particular broadcast station. The reduction in power reduces the strength of the 8VSB signal to some degree. The loss of 8VSB signal strength becomes objectionably large as the proportion of P-2VSB signal time-division multiplexed with the 8VSB signal is increased beyond a few percent.
U.S. patent application Ser. No. 10/078,933 teaches that trellis-coded 4VSB modulation can be generated by choosing the Y2 bit in each two-bit symbol supplied to a ⅔ trellis encoder in an 8VSB DTV transmitter to be the same as the Z0 bit generated by the ⅔ trellis encoder. This results in ⅓ trellis encoding that restricts the 8VSB symbol alphabet so as to consist of only those symbols corresponding to normalized modulation levels of −7, −3, +3 and +7. A symbol-error distance of at least four normalized modulation levels is required to cause a simple data slicer to make an error in decision as to the received Z1 information bit. Presuming the transmitted data to be random, the trellis-coded 4VSB modulation has a ratio of average power to peak power that is nominally the same as 8VSB modulation has.
U.S. patent application Ser. No. 10/078,933 describes P-2VSB and trellis-coded 4VSB modulation being generated subsequent to the convolutional interleaving of data segments that follows Reed-Solomon coding. Such procedure undesirably complicates the time-division multiplexing of P-2VSB and trellis-coded 4VSB signal with ordinary 8VSB signal. U.S. patent application Ser. No. 10/733,645 filed 12 Dec. 2003 by A. L. R. Limberg and titled “ROBUST SIGNAL TRANSMISSIONS IN DIGITAL TELEVISION BROADCASTING” was published 25 Nov. 2004 with publication number 2004-0237024 and issued 27 Mar. 2007 as U.S. Pat. No. 7,197,685. That published patent application describes P-2VSB being generated subsequent to data randomization, but prior to the convolutional interleaving of data segments that follows Reed-Solomon coding. Such procedure simplifies the time-division multiplexing of P-2VSB with ordinary 8VSB signal, since interleaving can be done on a data-segment-by-data-segment basis. This facilitates assembly of a program originating from more than one source. Switching among several TV cameras is simpler, and so is switching between live TV and recorded TV. Editing recorded TV is simpler.
U.S. patent application Ser. No. 10/955,212 filed 30 Sep. 2004 by A. L. R. Limberg and titled “TIME-DEPENDENT TRELLIS CODING FOR MORE ROBUST DIGITAL TELEVISION SIGNALS” was published 7 Apr. 2005 with publication number 2005-0074074. That publication in its entirety is incorporated herein by reference. That published patent application describes PCMP (prescribed coset pattern modulation) for DTV signals, describing specific types of PCPM as well as describing PCPM generically. In PCMP the 8VSB symbol alphabet is restricted in one of two prescribed ways for each symbol. The Y2 bit in each two-bit symbol supplied to the ⅔ trellis encoder in an 8VSB DTV transmitter is chosen in accordance with a prescribed value to restrict the symbol alphabet. If the prescribed Y2 bit is ONE, the symbol alphabet is restricted to a first coset of −3, −1, +5 and +7 normalized modulation levels. If the prescribed Y2 bit is ZERO, the symbol alphabet is restricted to a second coset of −7, −5, +1 and +3 normalized modulation levels. The pattern of prescribed Y2 bits is chosen so that the ratio of average power to peak power for PCPM is nominally the same as for 8VSB modulation. A symbol-error-distance of at least six normalized modulation levels is required to cause a simple data slicer to make an error in decision as to the received Z2 information bit.
Published patent application No. 2005-0074074 describes PCPM being generated subsequent to data randomization, but prior to the convolutional interleaving of data segments that follows Reed-Solomon coding. This procedure facilitates the time-division multiplexing of PCPM signal with ordinary 8VSB signal.
U.S. Pat. No. 6,178,209 titled “METHOD OF ESTIMATING TRELLIS ENCODED SYMBOLS UTILIZING SIMPLIFIED TRELLIS DECODING” issued 19 Jun. 1998 to S. N. Hulyalkar, T. J. Endres, T. A. Schaffer and C. H. Strolle. U.S. Pat. No. 6,178,209 describes a “smart” data slicer that takes into consideration the Z0 bit predicted by trellis code when making a symbol decision. The symbol-error-distance of 8VSB signal is at least two normalized modulation levels with a simple data slicer. Theoretically, a smart data slicer can double the symbol-error-distance of 8VSB signal to four normalized modulation levels. A smart data slicer can be modified such that in theory a symbol-error-distance of at least twelve normalized modulation levels is required to cause an error in decision as to the received Z1 information bit in a P-2VSB signal. A smart data slicer can be modified such that in theory a symbol-error-distance of at least eight normalized modulation levels is required to cause an error in decision as to the received Z2 information bit in a PCPM signal. A smart data slicer is not particularly advantageous when symbol decoding trellis-coded 4VSB modulation.
The entirety of U.S. patent application Ser. No. 11/119,662 titled “DIGITAL TELEVISION SIGNALS USING LINEAR BLOCK CODING” filed 2 May 2005 by A. L. R. Limberg is included herein by reference. Application Ser. No. 11/119,662 describes linear block coding of digital data to be transmitted using PCPM. This application, now abandoned, was published 2 Nov. 2006 with publication number 2006-0245505.